Two positions starting in early 2027 at LITIS / INSA Rouen Normandie

  1. MSc / final-year engineering internship — Graph Machine Learning The topic is: When is graph structure actually useful? The goal is to study whether the reconstructibility of graph edges from node features can explain when and why GNNs outperform feature-only models such as MLPs.

The internship will involve Graph ML, GNN/MLP comparisons, experimental evaluation, and statistical analysis. It is a 5–6 month internship starting from January 2027. Previous experience in Graph ML is welcome but not required for candidates with solid ML and Python foundations.

  1. Research Engineer — Machine Learning for polymer-property prediction This is a 6-month full-time position within the ANR OCTOPUSSY project, at the interface between machine learning, theoretical chemistry, and polymer science.

The work focuses on predicting polymer glass-transition temperature from 2D molecular graphs, conformer ensembles, quantum descriptors, and chemical language models. The project already includes an original dataset and an established experimental pipeline, with the objective of consolidating the benchmark and preparing a scientific publication.

Profiles from engineering school, MSc, or PhD level are welcome, with strong Python/PyTorch skills. Experience with GNNs, RDKit, Transformers, Slurm, or HPC would be a plus.

Both positions are based at LITIS, INSA Rouen Normandie, near Rouen, France.

Full details:

Applications and questions can be sent to: benoit.gauzere@insa-rouen.fr